Learning Forum

[felix512]

Hello, is there someone experimented in combustion simulations who can help me please ?

Simulation 1 : I made a methane - air combustion which worked perfectly. The geometry was with a bluff between the air and the fuel inlet.

Simulation 2 :  Same as Simulation 1, I just removed the bluff (so now I have 2 parallel pipes for the inlets), but now there is no combustion occuring.

Here are the parameters of the simulations :
I'm doing a steady 2D axisymmetric simulation, of methane - air combustion.
Using Species Transport model, with Gri-Mechanism and EDC
Standard k-e model for turbulence, and adiabatic walls.

For the simulation 1 with the bluff, I obtained an ignition after putting a wall on the bluff at 2000K. But when I remove the bluff (Simulation 2) , even setting a wall at 2000K between air inlet and fuel inlet doesn't work.

When I patch a high temperature, or species at 0.01 before or during the simulation, I still don't have ignition. Same if I put the air and fuel inlet at 1300K, the results aren't good.
When I use Eddy Dissipation model it also doesn't work.

Can someone help me please ?

I think that's a problem of turbulence. EDC needs a minimum of turbulences to starts chemical reactions. With the bluff geometry, enough turblences were created, but when I remove the bluff, there are maybe not enough turbulences to start the reactions ?

Thank you very much for helping.

[Pictures of Bluff and no bluff geometries] :

[Rob]

The mesh won't help. Please review both the resolution but also cell growth rates the AIS courses should be back up too. Also remember, if you don't have anything to stabilise the flame should there be any combustion if it's blown out? 

[felix512]

 

Thanks for your repley.

Indeed, the mesh I attached has a low quality. I already tried with high refined mesh, but it also doesn’t work. I went back to low mesh to improve calculation time, because some simulations were really really long (for example after patching temperature and/or species, the calculation time is extremly high, more than 20min for 1 iteration).

I understand that the problem might be blowoff limit of the flame, but the burner that I’m studying have a ‘simple’ geometry, without swirl/pre-heanting or other thing to stabilize the flame. A worker already did a thesis with this burner and succeed to obtain good stabilization with ‘real’ experiment. Hence, I think the problem might comes from my model…

I succeed to modelize using the Non-Premixed combustion, with Steady-Diffusion Flamelet + prompt and thermal NOx => perfect results.

My problem is that for the future, I would like to add ammonia in my model, and I’ll need fuel NOx calculations, which is impossible with Non-Premixed model without activating Discrate Phase, not suited for my case (I think).

 

Hence, the EDC model seems to be the best solution in my case, since this model can calculate NOx model with the mechanism imported.

 

Do you have any advice please ? Thanks.

 

[Rob]

If you're needing that long per iteration check what hardware you're using relative to the cell count. As you add physics you're unlikely to improve the model stability, and additional physics will also put the RAM and cpu load up. 

How is the ammonia added to the burner? Ie is it a droplet spray or do you add gas/vapour to the fuel or oxidant streams?

[felix512]

What do you mean exactly with add physics please ? I don't get it.

- I'll check the coherence between my hardware part and my simulation, but the strange fact is that time calculation seems normal on my computer, excepted after patching. Without patching, EDC calculations are fast it's ok.

- I'm adding gaseous ammonia (fuel inlet composed of gaseous CH4 / H2 / NH3). That's a coaxial burner with a very simple geometry. My EDC model works for 100% hydrogen fuel, but doesn't ignite for 100% methane; while the experiment did it with 100% methane in the burner. 

This information might be useful :  When I do 100% methane combustion using Non-Premixed model : it works. Then switching to EDC, with just species and energy equations solving, creates a detachment (the flame blows off)  and combustion stops.

Thanks for replying quickely !

[Rob]

By adding physics I mean more models, so higher order combustion or turbulence models. 

[Ren]

1. If the Eddy Dissipation Model (EDM) does not produce combustion, then two most likely reasons are:

flow is not turbulent. Please check turbulence levels (Re, or turbulent viscosity ratio).

Not all reactant species mass fractions are nonzero, and at least one product species mass fraction is nonzero at the beginning of the iteration. Please patch some nonzero values.

2. The EDC model requires adequate turbulence and artificial ignition. It's better to start the EDC calculaiton from a reasonably converged EDM solution or solution from "Relax to Chemical Eqilibrium" calculation.